Background: Changes of toxic metals and essential elements during childhood may be the risk factor of autism spectrum disorder (ASD). This research established an accurate personalized predictive model of ASD behaviors among children by using the blood element detection index of children in Xinjiang, China.
Methods: A total of 1537 children (240 ASD behavior children and 1297 non-ASD behavior children) aged 0-7 were collected from September 2018 to September 2019 in Urumqi Children's Hospital and the health management institute of Xinjiang Medical University. For measuring the copper (Cu), zinc (Zn), magnesium (Mg), iron (Fe), calcium (Ca), lead (Pb), and cadmium (Cd), 80 μL of blood was taken from each participant's ring finger. Univariate logistic regression analysis was used to select predictors, then the multivariate logistic regression was used to establish the predictive model. The discriminability, calibration and clinical validity of the model were evaluated by the receiver operating characteristic (ROC) curve, Hosmer-Lemeshow test and decision curve analysis (DCA).
Results: Gender, concentrations of Pb, Ca and Zn in children's blood specimens were found to be the independent risk factors of ASD behaviors and were used to develop the nomogram model. The area under the ROC curve (AUC) in the development group (AUC = 0.778) and the validation group (AUC = 0.775) showed the model had discrimination ability. The calibration curve indicated the model was accurate, and the DCA proved its clinical application value.
Conclusion: The nomogram model can be used as a reliable tool to predict the risk of ASD behaviors among children.
Keywords: Autism spectrum disorder (ASD); Blood elements; Nomogram; Predictive model.
Copyright © 2022 Elsevier B.V. All rights reserved.